FIGURE LEGENDS
Figure 1. Worldwide map showing in colours the sampling locations for the specimens used in RADseq analyses. Dates in parenthesis represent the earliest record based on literature and the Natural History Museum of London collection (e.g. St Helena Island). Dark grey countries: Colombia (CO), Israel (IS) and the asterisk (*) indicated medflies collected for microbiome analyses; the numbers indicate the host plant (note that some flies from the same sampling site were collected in different host plants).
Figure 2. Individual genotype assignment for the six populations of medfly distributed worldwide. A) DAPC plot for the clustering obtained in AIC analysis (K = 3). B) Individual genotype assignment inferred by STRUCTURE with K =2 and K=3 top and bottom, respectively.
Figure 3. Individual genotype assignment for the locations in the introduced range of medfly. A) DAPC analysis for the clustering obtained in AIC analysis (K = 3) of 82 individuals. B) STRUCTURE results based on K =2 and K =3 top and bottom, respectively.
Figure 4. Simple co-ancestry heatmap for medfly populations. Top: the raw data matrix tree (simple hierarchical “tree-like” with posterior population assignment probabilities), each tip corresponds to a sampling site. Left: the location abbreviations. Bottom: the clusters identified in the co-ancestry matrix.
Figure 5. Best fit hypothetical evolutionary scenarios of the invasion routes of C. capitata. Left: topology of the best fit scenario selected from six different invasion routes (Scenario 1); South Africa was predicted as the ancestral population followed by colonisation of Brazil, and from there was predicted a divergence to the cluster Spain-Guatemala and Greece-Australia. Right: topology of the best-predicted scenario from three different invasion routes tested (Scenario 4). South Africa was the predicted ancestral population for medfly followed by divergence to Brazil. The clusters Spain-Guatemala and Greece-Australia appeared to be the result of admixture between South Africa and Brazil populations (see Methods and Supplementary Fig.S1 for details). Colours correspond to sampling sites codes in fig 2A (except sampling sites grouped).
Figure 6. Species tree of 30 individuals collected across six medfly sampling sites (SA: South Africa, BR: Brazil, SP: Spain, GU: Guatemala, AU: Australia). A total of 15 consensus tress topologies were obtained, here is represented the consensus tree 1 which covers 37.18% of the total cumulative trees. Branch width is proportional to theta represented in black.
Figure 7. Relative abundance of the different microorganisms by phyla detected across six sampling localities of medfly. Phyla with abundance lower than 1﹪are grouped into “Other”. On top Barr plots are the abbreviations to each population studied, each bar represented one specimen per sampling site, for more information refer to Table 1 and Supplementary Table 1.
Figure 8. Differential abundance of medfly microbiome at genus taxonomic level in Brazil (BR) and South Africa (SA). On the left, levels represent 48 ASVs with relative abundance above 0.01% and FDR under 0.01. Circles are colour coded according to bacterial phylum and RA stands for relative abundance. The ASV taxonomically assigned to Burkholderiaceae_unclassified is highlighted in blue to indicate its unique presence in all samples from Brazil.